Vision-based multisensor integration in remote-brained robots

This paper presents a method for the integration of multiple sensors in a "remote-brained robot", that is, one where processing capability is accessed by wireless links. The remote-brained approach allowed the authors to build a robot with a free body and a heavy brain. In order to integrate multiple sensors in a remote-brained robot, it is crucial to consider how to multiplex multisensor information. The authors have examined several configurations and developed a method using a video screen to integrate touch, force and visual information in a sensor image. This paper describes the possible configurations for multisensor integration in remote-brained robots and presents real examples of vision-based sensor integration for a dynamic biped robot and a multi-limbed apelike robot.<>